The proposed effort will develop a low cost, ultra-high resolution, small animal PET imaging system. The key component of the R21 phase of this proposal is the development and feasibility testing of a continuous miniature crystal element (cMiCE) detector design. The goals are a detector intrinsic spatial resolution that will translate into a camera image resolution of about 1 mm full width at half maximum (FWHM) and an effective detector imaging area greater than or equal to the imaging area of a discrete crystal detector using the same photomultiplier tube (PMT). The cost of ultra-high resolution detectors using 1x1 mm or smaller cross-section crystals is driven by the cost to cut and surface treat the individual crystals and to hand assemble the modules. By using a single 25x25 mm cross-section crystal, major cost reductions will be realized versus using an equivalent array of 1x1 mm crystals. A continuous crystal detector will also have better packing fraction and sampling characteristics. The main limitation to a small area continuous crystal implementation has been a reduced effective imaging area due to edge effects. A statistically based positioning technique is proposed to extend the imaging area of the crystal. A second drawback is that intrinsic spatial resolution tends to broaden as the crystal thickness increases. Techniques to improve the spatial resolution characteristics of thick continuous crystal detectors will also be investigated. Detector modules will use currently available scintillator materials and PMTs; however, the design will be easily adaptable to new technologies, for example, avalanche photodiode arrays if they become cost effective. Careful characterization of the detector performance will be undertaken. Issues such as detector stability, detector calibration, and count rate performance will be included. The R33 phase of the proposal will include building a dedicated small animal PET system using the cMiCE detector design. The preferred system design will have detector ring diameter of about 12 cm and an axial field of view approaching 10 cm. Monte Carlo simulation will be used to optimize the resolution-sensitivity (i.e., contrast versus noise) trade-off for estimation task performance. Additional simulations to quantify the effects of positron range for different radioisotopes will be conducted. The quantitative accuracy of images reconstructed with and without resolution recovery (e.g., positron range, photon acollinearity and detector response) will be studied. Finally, the feasibility of using the variation in the LRF to estimate depth of interaction for thick crystal designs will be investigated.